2
\$\begingroup\$

I am wondering is it possible to design a differential (or instrumentation) amplifier that will operate from single supply?

I have a 2Vp-p and 0.2Vp-p sine wave signals that need to go thru the differential amplifier. They both swing positive and negative.

But currently for this project I have only positive power supply. I can add an inverting converter if necessary but would like to know if I can reduce the costs.

If it is possible, could you please show an example.

Thanks

\$\endgroup\$
  • \$\begingroup\$ If your input signal has no dc component, you could re-bias it with capacitors or a transformer. Of course if those solutions work, you might also not need an in-amp to begin with. \$\endgroup\$ – The Photon Mar 10 '15 at 16:23
3
\$\begingroup\$

Getting the input of a amplifier to work outside its supply voltage range is difficult, but there are possible ways around this:

  1. Is the signal AC? If so, capacitively couple it to something that floats around ½ the supply voltage.

  2. Is the signal floating (not ground-referenced)? This can be the case when coming directly from some tranducer, for example. In that case, tie the other end to something around ½ the supply voltage.

  3. Shift the voltage level. This could be done with a resistor divider to the positive supply, for example. That will also attenuate the signal a little, but that can be made up by increasing the gain a little. Alternatively, the signal can be shifted by using zener diodes or ordinary diodes forward biased.

    Both these schemes will add some offset error. Everything is a tradeoff.

  4. make a negative supply. It's not really that hard or expensive, and there are off the shelf chips just for this purpose. For just powering a opamp, this could be a cpacitive charge pump, which will generally be cheaper than a inductive switcher.

    If you already have something that produces a regular clock, a capacitive charge pump can be as easy as a couple of diodes and caps. I have a few times used the clock output of a microcontroller for this purspose, although then I buffered it with two transistors making a push-pull emitter follower.

\$\endgroup\$
  • \$\begingroup\$ +1 for "make a negative supply" - even without being clever and making a charge pump yourself, a charge pump IC is not many components or dollars. \$\endgroup\$ – Greg d'Eon Mar 10 '15 at 17:16
  • \$\begingroup\$ Yes the signal is AC and it is not floating. I have to later pass this signal into an microcontroller. Would it be best to power the op-amps from dual supply and then shift the voltage up. Or power the op-amps from single supply and provide the correct biasing? Could you provide an example with your first point? \$\endgroup\$ – Golaž Mar 10 '15 at 18:14
  • \$\begingroup\$ @Gola: Make 1/2 the supply with two resistors, like 1 MOhm. Tie that to the amplifier input, and capacitively couple your signal onto that node also. The capacitor will pass your AC signal, but block DC, allowing both sides of the capacitor to be at different average voltages. \$\endgroup\$ – Olin Lathrop Mar 10 '15 at 18:48
  • \$\begingroup\$ @OlinLathrop But if I bias the voltages at the inputs at 1/2 the rail (lets say 2.5V) the voltages will be 1.5V to 3.5V and 2.4V to 2.6V (and lets say they are in phase). And if we for example subctract the peak voltages, we would get 3.5-2.6=0.9V and 1.5-2.4=-0.9V so we would get negative voltages. How can I fix this problem? \$\endgroup\$ – Golaž Mar 11 '15 at 17:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.